mlabonne

/

UltraMerge-7B

like 11

Hi, this is a fine-tune of automerger/YamShadow-7B that is still W.I.P because it doesn't reach the level of performance that I wanted. I don't think that reporting it is good practice here.

It's score is higher than NeuralHermes-2.5-Mistral-7B, share I change the model? hmmm.........